Instructions to use RESMPDEV/Mistral-7B-v0.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RESMPDEV/Mistral-7B-v0.2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="RESMPDEV/Mistral-7B-v0.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RESMPDEV/Mistral-7B-v0.2")
model = AutoModelForCausalLM.from_pretrained("RESMPDEV/Mistral-7B-v0.2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use RESMPDEV/Mistral-7B-v0.2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RESMPDEV/Mistral-7B-v0.2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RESMPDEV/Mistral-7B-v0.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RESMPDEV/Mistral-7B-v0.2

SGLang

How to use RESMPDEV/Mistral-7B-v0.2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "RESMPDEV/Mistral-7B-v0.2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RESMPDEV/Mistral-7B-v0.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "RESMPDEV/Mistral-7B-v0.2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RESMPDEV/Mistral-7B-v0.2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RESMPDEV/Mistral-7B-v0.2 with Docker Model Runner:
```
docker model run hf.co/RESMPDEV/Mistral-7B-v0.2
```

This is a hacked together version of the new Mistral-7b-v0.2 FP16 weights directly downloaded from their CDN.

The conversion was done by directly converting the monolithic pickle file to safetensors and building the index which is suboptimal.

Credit to Mistral AI and the amazing team over there and Cognitive Computations especially Eric Hartford for tutelage and helping navigate the LLM landscape.

~~As this is a mix of Mistral 7b v0.1 and Mistral 7b v0.2 files it is to be considered a pre-alpha.~~

This conversion is suboptimal and I would use https://huggingface.co/alpindale/Mistral-7B-v0.2-hf for the FP-16 weights until MistralAI does their offical release.

Downloads last month: 842

Safetensors

Model size

7B params

Tensor type

BF16